Trends in Alcohol Use Disorder Treatment Utilization and Setting From 2008 to 2017
Abstract
Objective:
Little is known about recent trends in treatment for alcohol use disorder. The authors used national data to examine treatment trends among individuals with alcohol use disorder.
Methods:
A sample of nonelderly adults (ages 18–64 years, N=36,707) with alcohol use disorder was identified from the National Survey on Drug Use and Health. Multinomial logistic regression analysis was conducted to examine trends in treatment for alcohol use disorder in 2008–2010, 2011–2013, and 2014–2017 in any medical setting (hospitals, rehabilitation centers, mental health centers, emergency departments, and private doctors’ offices), self-help groups only (no medical setting), and no setting (i.e., no treatment). Additional analyses investigated trends in mental health treatment. Regression models adjusted for predisposing, enabling, and need-related characteristics.
Results:
Among those with an alcohol use disorder, the percentage who received any treatment was significantly lower in 2011–2013 (5.6%) than in 2008–2010 (6.9%) (p<0.05). In adjusted analyses, the probability of receiving no treatment increased by 1.5 percentage points in 2014–2017 (95% CI=0.5–2.5) compared with the 2008–2010 baseline. Significant declines were observed in the receipt of any treatment in a medical setting (marginal effect [ME]=−1.0%, 95% CI=−2.0 to −0.0) and self-help treatment only (ME=−0.5%, 95% CI=−0.8 to −0.1) in 2014–2017 compared with the baseline period. The probability of receiving any mental health treatment did not change during the study period.
Conclusions:
Among persons with an alcohol use disorder, treatment declined from 2008 to 2017. Future studies should examine the mechanisms that may be responsible for this decline.
HIGHLIGHTS
•
Data from the 2008–2017 National Survey on Drug Use and Health were used to study trends in treatment utilization and changes in treatment setting among persons with alcohol use disorder.
•
Alcohol use disorder treatment declined by 1.1 percentage points over this period, from 6.9% in 2008–2010 to 5.8% in 2014–2017.
•
In adjusted models, compared with the 2008–2010 period, the probability of receiving no treatment increased by 1.4 percentage points in 2011–2013 and 1.5 percentage points in 2014-2017.
•
Compared with the 2008–2010 period, the likelihood of receiving treatment in any medical setting declined by 1.0 percentage points and of receiving only self-help treatment by 0.5 percentage points in 2014–2017.
As of 2018, alcohol use disorder was the most common substance use disorder in the United States, affecting 5.4% of Americans ages >12 years (1). Alcohol use disorder is associated with various medical problems, including cancer, diabetes, and cardiovascular disease, and from 2011 to 2015, it was associated with 95,158 deaths a year in the United States (2–5). Although effective treatments for alcohol use disorder exist, only about 10% of those with alcohol use disorder receive treatment in a given year, and many individuals with alcohol use disorder seek mental health treatment instead of specific treatment for this disorder (6, 7).
In the past decade, two significant laws were implemented that had the potential to address the treatment gap for alcohol use and other behavioral health disorders—the Mental Health Parity and Addiction Equity Act (MHPAEA) in 2008 and the Patient Protection and Affordable Care Act (ACA) in 2010. These laws increased the number of insured individuals, increased coverage for behavioral health treatment among the insured, and more tightly integrated behavioral health care and medical health care (8–13). The MHPAEA established parity for substance use disorder benefits, requiring that both treatment limits (e.g., caps on outpatient visits) and financing (e.g., copays) be no more restrictive for substance use treatment than for medical benefits (8). The ACA extended parity to individual and Medicaid plans and required coverage of substance use disorder benefits as a part of the essential health benefits (9). The ACA also increased the number of insured individuals through the individual marketplace and Medicaid expansion (10, 12). Finally, payment reform and other ACA incentives were expected to lead to greater integration of substance use disorder treatment with primary care, facilitating access to treatment (11, 13).
Several studies have examined treatment trends for substance use disorders since the passages of the MHPAEA and ACA (14–18). These studies have typically found no change in the treatment rate for substance use disorders. However, the studies have aggregated treatment for all substance use disorders, potentially masking differential trends for illicit drug use disorders and alcohol use disorder. To date, only one known study has used national data (from 2010 to 2015) to specifically investigate recent trends in alcohol use disorder treatment (19). This study reported no change in alcohol use disorder treatment over that period. Because it included data only through 2015 (the first full year after ACA implementation), more recent data are needed to understand how trends in treatment for alcohol use disorder have changed since the ACA’s implementation. In addition, previous research examined trends in alcohol use disorder treatment across all settings in aggregate (including medical settings and self-help groups) rather than by treatment setting.
To address this literature gap, we used national data from the 2008–2017 National Survey on Drug Use and Health (NSDUH) to examine trends in treatment for alcohol use disorder. We hypothesized that alcohol use disorder treatment increased over that period, in part because of the increase in insurance coverage resulting from the ACA. We examined three measures of alcohol use disorder treatment (no treatment, any treatment in a medical setting, and self-help treatment only) and mental health treatment for those with alcohol use disorder.
Methods
Data and Sample
We used pooled data from the 2008–2017 NSDUH. The NSDUH is an annual, nationally representative, cross-sectional survey of the US civilian, noninstitutionalized population ages ≥12 years (20). It uses DSM-IV criteria to determine whether respondents have an alcohol use disorder (i.e., alcohol abuse or dependence). We identified 38,335 individuals with alcohol use disorder ages 18–64 years in the pooled sample, of whom 36,707 had complete information on all of the model covariates for analyses to examine the three alcohol use disorder treatment measures. An additional 79 observations were missing on the measure of mental health treatment and were removed from the analytic sample used in a separate mental health treatment model for those with alcohol use disorder. Similar to previous research examining behavioral health services changes over time by using NSDUH data, we examined trends in alcohol use disorder treatment across three periods: 2008–2010, 2011–2013, and 2014–2017 (14, 16, 18, 19). This study was exempt from institutional review board review in a letter of determination from Emory University’s Institutional Review Board because the data came from publicly available sources without private health information identifiers.
Dependent Variables
We measured the receipt of any alcohol use disorder treatment by using a categorical variable (received no treatment, received any medical treatment, or received only self-help treatment). We derived this measure from a question that asked respondents whether they had received treatment for alcohol or drug use in the past 12 months and whether that treatment was for use of alcohol, drugs, or both. Those who did not report treatment for alcohol use or both alcohol and drug use were coded as not receiving treatment for alcohol use.
Respondents who received treatment were categorized on the basis of where they reported receiving treatment—that is, whether they received it in a medical setting (i.e., as a hospital inpatient, in a rehabilitation center, mental health center, emergency room, or private doctor’s office) or whether they received only self-help treatment, such as Alcoholics Anonymous. We excluded from analyses those who received treatment only in jail or for whom treatment setting could not be ascertained (N=654). Consistent with previous studies using NSDUH (21–23), we aggregated treatment settings into these categories to separate settings where health insurance coverage was likely to facilitate financial access to services (medical settings) from those in which insurance was unlikely to make a difference (self-help treatment).
The receipt of mental health treatment was measured as a dichotomous variable (yes or no). Respondents who reported receiving inpatient treatment, outpatient treatment, or prescription medication for mental health treatment were coded as receiving mental health treatment. We included mental health treatment as a separate dependent variable because previous research has found that individuals with alcohol use disorder are more likely to seek mental health treatment than substance use disorder treatment and that the overall number of those receiving outpatient mental health treatment had increased in the years after ACA passage (7, 24). However, it is unclear to what extent utilization of mental health treatment changed for those with alcohol use disorder.
Independent Variables
Sociodemographic characteristics.
Regression models controlled for predisposing and enabling characteristics (25). Predisposing characteristics included age category (18–25, 26–34, 35–49, and 50–64 years), race-ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and non-Hispanic other), gender (male and female), educational attainment (less than high school, high school graduate, some college or associate’s degree, and college graduate and above), marital status (married, widowed or divorced, and never married), and employment status (full-time employed, part-time employed, unemployed, and other). Because of small sample sizes, we combined four race-ethnicity categories into the non-Hispanic other category for analysis: non-Hispanic Native American–Alaska Native, non-Hispanic Native Hawaiian–other Pacific Islander, non-Hispanic Asian, and non-Hispanic more than one race. Enabling characteristics included income (<$20,000, $20,000–$49,999, $50,000–$74,999, and ≥$75,000), and insurance status (any private insurance, Medicaid [no private insurance], other insurance [no Medicaid or private insurance and other insurance], and uninsured). Income was imputed for 8.6% (N=3,167) of the sample.
Health status.
Regression models also adjusted for differences in need-related characteristics (25). Health status measures included a categorical measure of self-reported health status (excellent, very good, good, fair, and poor) and dichotomous indicators for comorbid illicit substance use disorder and comorbid mental disorder, as well as alcohol use disorder severity. Comorbid substance use disorder included heroin abuse or dependence, cocaine abuse or dependence, and marijuana abuse or dependence; these measures of illicit substance use or dependence were the only measures consistently assessed during the study period (26). The measure of comorbid mental health status was created with information from the NSDUH prediction model that determines the predicted probability of any mental illness on the basis of responses to several questions; individuals with a predicted probability greater than or equal to a specified cutoff value were considered to have a comorbid mental disorder (26). Alcohol use disorder severity was assessed with an indicator for whether the individual met DSM-IV criteria for alcohol dependence (vs. alcohol abuse). In addition, the models included indicators for the following self-reported health conditions: diabetes, liver cirrhosis, hepatitis, asthma, HIV or AIDS, and high blood pressure. Each condition was assessed with an indicator for whether the individual reported ever being told by a physician or other health professional that they had each respective condition.
Data Analysis
Analyses were conducted in Stata, version 16, by using the SVY command to account for the NSDUH’s complex survey design elements. We first conducted bivariate analyses with adjusted Wald tests to compare treatment rates in 2011–2013 and 2014–2017 with those in the baseline period of 2008–2010. Next, we adjusted for predisposing, enabling, and need-related factors that may have changed over time by using multinomial and binomial logistic regression models.
We estimated a multinomial logistic regression for the dependent variable of alcohol treatment: received no treatment, received any treatment in a medical setting, and received only self-help group treatment (i.e., no medical treatment). We estimated a separate logistic regression for the dependent variable of any mental health treatment. Marginal effects were reported for all models and can be interpreted as the percentage-point change in the predicted probability that individuals fall into one outcome category compared with the other outcome categories combined, holding all other covariates constant. The alpha level for statistical significance was set to 0.05.
Results
The prevalence of alcohol use disorder among all NSUDH respondents in our sample declined from 2008 to 2017. Bivariate analysis indicated a decline in this disorder from 8.9% in 2008–2010 to 8.1% in 2011–2013 (p<0.001) and 7.3% in 2014–2017 (p<0.001) (Table 1). However, among those with alcohol use disorder, a higher proportion had alcohol dependence, the more severe form of alcohol use disorder, in 2014–2017 (52.3%) than in 2008–2010 (46.8%, p<0.001) (Table 2).
| Period | N | Weighted % |
|---|---|---|
| 2008–2010 (N=107,893) | 13,432 | 8.9 |
| 2011–2013 (N=107,713) | 11,733 | 8.1a |
| 2014–2017 (N=155,868) | 13,170 | 7.3a |
a
p<0.001; adjusted Wald test comparing 2011–2013 and 2014–2017 values with the 2008–2010 baseline period.
| 2008–2010 (N=12,810b/12,788c) | 2011–2013 (N=11,211b/11,186c) | 2014–2017 (N=12,686b/12,654c) | ||||||
|---|---|---|---|---|---|---|---|---|
| Variable | N | Weighted % | N | Weighted % | p | N | Weighted % | p |
| Alcohol treatment | ||||||||
| Any | 780 | 6.9 | 573 | 5.6 | .033 | 738 | 5.8 | .054 |
| Medical setting | 644 | 5.6 | 454 | 4.4 | .030 | 626 | 5.0 | .252 |
| Self-help only | 136 | 1.2 | 119 | 1.1 | .684 | 112 | .8 | .024 |
| Mental health treatment | 2,504 | 22.6 | 2,373 | 23.3 | .365 | 2,922 | 23.9 | .130 |
| Age in years | ||||||||
| 18–25 | 8,731 | 32.7 | 7,415 | 30.5 | .019 | 5,756 | 26.6 | <.001 |
| 26–34 | 1,878 | 25.3 | 1,650 | 25.4 | .920 | 3,042 | 25.2 | .918 |
| 35–49 | 1,722 | 26.9 | 1,598 | 27.6 | .569 | 2,933 | 27.4 | .674 |
| 50–64 | 479 | 15.1 | 548 | 16.4 | .237 | 955 | 20.8 | <.001 |
| Male | 7,749 | 65.9 | 6,658 | 64.0 | .043 | 7,456 | 63.2 | .009 |
| Race-ethnicity | ||||||||
| Non-Hispanic White | 8,602 | 69.6 | 7,274 | 67.9 | .115 | 8,002 | 65.6 | <.001 |
| Non-Hispanic Black | 1,218 | 11.0 | 1,125 | 10.2 | .377 | 1,318 | 11.1 | .844 |
| Hispanic | 1,910 | 15.2 | 1,792 | 16.7 | .049 | 2,074 | 16.7 | .043 |
| Non-Hispanic other | 1,080 | 4.2 | 1,020 | 5.2 | .014 | 1,292 | 6.6 | <.001 |
| Marital status | ||||||||
| Married | 2,181 | 32.3 | 1,879 | 32.3 | .969 | 3,053 | 33.2 | .437 |
| Widowed or divorced | 1,121 | 16.1 | 1,044 | 16.3 | .829 | 1,551 | 16.4 | .737 |
| Never married | 9,508 | 51.6 | 8,288 | 51.5 | .906 | 8,082 | 50.4 | .285 |
| Income in $ | ||||||||
| <20,000 | 3,810 | 22.2 | 3,477 | 23.4 | .234 | 3,086 | 20.5 | .047 |
| 20,000–49,999 | 4,189 | 31.0 | 3,691 | 31.5 | .619 | 3,973 | 28.3 | .002 |
| 50,000–74,999 | 1,908 | 16.8 | 1,505 | 14.8 | .031 | 1,893 | 15.2 | .044 |
| ≥75,000 | 2,903 | 30.0 | 2,538 | 30.3 | .863 | 3,734 | 35.9 | <.001 |
| Self-reported health status | ||||||||
| Excellent | 2,871 | 20.3 | 2,404 | 20.2 | .877 | 2,444 | 18.5 | .012 |
| Very good | 5,249 | 39.3 | 4,626 | 38.5 | .553 | 5,044 | 38.5 | .447 |
| Good | 3,523 | 28.9 | 3,084 | 28.2 | .475 | 3,749 | 29.8 | .293 |
| Fair | 1,028 | 9.6 | 933 | 10.6 | .091 | 1,273 | 11.0 | .041 |
| Poor | 139 | 1.9 | 164 | 2.5 | .147 | 176 | 2.1 | .506 |
| Education | ||||||||
| Less than high school | 2,191 | 15.8 | 1,690 | 14.1 | .050 | 1,537 | 11.7 | <.001 |
| High school graduate | 4,021 | 29.8 | 3,366 | 27.2 | .025 | 3,280 | 24.3 | <.001 |
| Some college or associate’s degree | 4,202 | 29.9 | 3,868 | 30.8 | .362 | 4,569 | 33.4 | <.001 |
| College graduate and above | 2,396 | 24.5 | 2,287 | 27.9 | .001 | 3,300 | 30.6 | <.001 |
| Employment status | ||||||||
| Full-time | 6,425 | 58.7 | 5,633 | 57.9 | .514 | 7,442 | 61.3 | .021 |
| Part-time | 2,753 | 16.9 | 2,410 | 15.9 | .244 | 2,199 | 15.1 | .013 |
| Unemployed | 1,514 | 9.9 | 1,286 | 9.3 | .337 | 1,089 | 7.6 | <.001 |
| Other | 2,118 | 14.5 | 1,882 | 17.0 | .005 | 1,956 | 16.0 | .023 |
| Comorbid substance use disorder | 1,997 | 11.4 | 1,668 | 11.1 | .664 | 1,685 | 11.3 | .868 |
| Comorbid mental illness | 4,458 | 36.0 | 4,114 | 38.8 | .009 | 5,282 | 39.3 | .001 |
| Alcohol dependence | 5,635 | 46.8 | 5,123 | 47.1 | .790 | 6,396 | 52.3 | <.001 |
| N of alcohol use disorder criteria | ||||||||
| 1–2 | 4,224 | 33.6 | 3,633 | 33.3 | .840 | 3,908 | 30.7 | .003 |
| 3–4 | 5,362 | 40.6 | 4,816 | 41.3 | .550 | 5,420 | 42.0 | .101 |
| ≥5 | 3,224 | 25.8 | 2,762 | 25.4 | .649 | 3,358 | 27.3 | .061 |
| Insurance status | ||||||||
| Private | 7,445 | 61.4 | 6,681 | 60.6 | .558 | 8,009 | 64.5 | .004 |
| Medicaid | 1,189 | 8.0 | 1,127 | 8.3 | .585 | 1,837 | 13.6 | <.001 |
| Other insured | 774 | 5.9 | 769 | 6.0 | .746 | 871 | 6.3 | .424 |
| Uninsured | 3,402 | 24.8 | 2,634 | 25.0 | .824 | 1,969 | 15.6 | <.001 |
| Health indicators | ||||||||
| Diabetes | 224 | 3.1 | 223 | 3.3 | .690 | 405 | 4.4 | .007 |
| Cirrhosis | 23 | .3 | 29 | .4 | .451 | 45 | .5 | .217 |
| Hepatitis B or C | 96 | 1.3 | 94 | 1.7 | .121 | 157 | 1.9 | .028 |
| Asthma | 1,852 | 13.0 | 1,607 | 13.4 | .535 | 1,607 | 11.7 | .057 |
| HIV or AIDS | 32 | .4 | 22 | .2 | .340 | 40 | .4 | .801 |
| High blood pressure | 1,113 | 15.2 | 1,080 | 15.8 | .509 | 1,234 | 13.7 | .048 |
a
p values from adjusted Wald test comparing 2011–2013 and 2014–2017 values with the 2008–2010 baseline period.
b
Sample for alcohol treatment model.
c
Sample for mental health treatment model.
Among those with an alcohol use disorder, bivariate analysis indicated that an overall decline in any treatment for alcohol use from 6.9% in 2008–2010 to 5.6% (p<0.05) in 2011–2013 (Table 2). We also noted a significant decline in treatment in any medical setting from 5.6% in 2008–2010 to 4.4% in 2011–2013 (p<0.05). Both any treatment and treatment in any medical setting had nonsignificant declines from 2008–2010 to 2014–2017. Self-help treatment exhibited only a small decline, from 1.2% in 2008–2010 to 0.8% in 2014–2017 (p<0.05). Mental health treatment increased nonsignificantly, from 22.6% in 2008–2010 to 23.9% in 2014–2017.
Next, we examined differences in characteristics of those with an alcohol use disorder across these periods (Table 2). Notably, individuals with alcohol use disorder had significant changes in health insurance status. A higher percentage of those with alcohol use disorder reported having Medicaid coverage (13.6%) and private insurance (64.5%) in 2014–2017 compared with the 2008–2010 period (8.0% and 61.4%, respectively, p<0.01). The percentage of those with alcohol use disorder who were uninsured declined from 24.8% in 2008–2010 to 15.6% in 2014–2017 (p<0.001). Other enabling characteristics also showed significant changes. A higher percentage of those with alcohol use disorder had an income ≥$75,000 per year (35.9%) and had a college degree or higher (30.6%) in 2014–2017 compared with 2008–2010 (30.0% and 24.5%, respectively, p<0.001).
We estimated a multinomial logistic regression to assess whether there were significant changes in the probability of treatment for alcohol use disorder, after controlling for changes in predisposing, enabling, and need characteristics during the study period (Table 3). Compared with the 2008–2010 period, the probability of not receiving treatment (vs. receiving treatment in any setting) increased by 1.4 percentage points (95% CI=0.4–2.4) in 2011–2013 and 1.5 percentage points (95% CI=0.5–2.5) in 2014–2017, after adjustments for confounders. Treatment in medical settings and self-help treatment only also significantly declined. Compared with the baseline period, the probability of receiving treatment in any medical setting declined by 1.3 percentage points in 2011–2013 (95% CI=−2.2 to −0.4) and 1.0 percentage points in 2014–2017 (95% CI=−2.0 to −0.0). Moreover, the probability of receiving self-help treatment declined only by 0.5 percentage points in 2014–2017 (95% CI=−0.8 to −0.1).
| No treatment (baseline=93.9%) | Any medical setting (baseline=5.1%) | Self-help only (baseline=1.1%) | ||||
|---|---|---|---|---|---|---|
| Variable | ME | 95% CI | ME | 95% CI | ME | 95% CI |
| Period (reference: 2008–2010) | ||||||
| 2011–2013 | 1.4** | .4, 2.4 | –1.3** | –2.2, −.4 | −.1 | −.5, .3 |
| 2014–2017 | 1.5** | .5, 2.5 | –1.0* | –2.0, −.0 | −.5** | −.8, −.1 |
| Male | –2.0** | –2.8, –1.1 | 1.4** | .7, 2.2 | .5** | .2, .9 |
| Age in years (reference: 18–25) | ||||||
| 26–34 | –2.1** | –3.3, −.8 | 1.9** | .7, 3.2 | .1 | −.3, .5 |
| 35–49 | –4.2** | –5.8, –2.6 | 4.1** | 2.6, 5.6 | .1 | −.4, .5 |
| 50–64 | –3.1* | –5.8, −.3 | 3.4* | .7, 6.1 | −.4 | −.9, .2 |
| Race-ethnicity (reference: non-Hispanic White) | ||||||
| Non-Hispanic Black | 1.9** | .9, 2.9 | –1.4** | –2.4, −.5 | −.5** | −.8, −.1 |
| Hispanic | 1.7** | .7, 2.7 | –1.8** | –2.7, −.9 | .1 | −.4, .5 |
| Non-Hispanic other | 1.5* | .3, 2.7 | –1.3* | –2.3, −.2 | −.2 | −.7, .3 |
| Marital status (reference: married) | ||||||
| Widowed or divorced | –2.0** | –3.5, −.5 | 2.1** | .7, 3.5 | −.1 | −.6, .4 |
| Never married | –1.3* | –2.6, −.1 | 1.5* | .3, 2.7 | −.1 | −.6, .3 |
| Income in $ (reference: ≥75,000) | ||||||
| <20,000 | –1.3 | –2.9, .2 | 1.5* | .0, 3.0 | −.2 | −.6, .2 |
| 20,000–49,999 | −.6 | –2.0, .9 | .7 | −.6, 2.1 | −.2 | −.6, .3 |
| 50,000–74,999 | –1.2 | –3.1, .7 | 1.3 | −.5, 3.2 | −.1 | −.6, .3 |
| Self-reported health status (reference: excellent) | ||||||
| Very good | −.3 | –1.6, 1.1 | .3 | –1.1, 1.6 | .0 | −.5, .5 |
| Good | –1.4 | –2.9, .1 | 1.3 | −.1, 2.8 | .1 | −.5, .6 |
| Fair | .2 | –1.3, 1.8 | −.7 | –2.1, .8 | .4 | −.4, 1.3 |
| Poor | .8 | –1.5, 3.1 | −.3 | –2.5, 1.9 | −.5 | –1.5, .4 |
| Insurance status (reference: uninsured) | ||||||
| Private | −.5 | –1.7, .6 | .7 | −.4, 1.7 | −.1 | −.6, .3 |
| Medicaid | –2.2* | –3.9, −.4 | 2.3* | .5, 4.0 | −.1 | −.5, .4 |
| Other insured | –1.5 | –3.3, .3 | .9 | −.5, 2.4 | .5 | −.5, 1.6 |
| Education (reference: high school graduate) | ||||||
| Less than high school | −.3 | –1.4, .8 | .1 | −.9, 1.1 | .2 | −.5, .9 |
| Some college | −.7 | –2.0, .5 | .2 | −.8, 1.3 | .5 | −.0, 1.0 |
| College graduate and above | .2 | –1.3, 1.7 | −.4 | –1.9, 1.0 | .2 | −.3, .8 |
| Employment status (reference: full-time) | ||||||
| Part-time | −.9 | –2.3, .5 | 1.1 | −.3, 2.4 | −.2 | −.7, .3 |
| Unemployed | –3.6** | –5.4, –1.9 | 3.8** | 2.1, 5.5 | −.1 | −.8, .5 |
| Other | –2.8** | –4.4, –1.2 | 3.0** | 1.5, 4.5 | −.2 | −.7, .3 |
| Comorbid substance use disorder | –5.1** | –6.5, –3.8 | 4.4** | 3.1, 5.6 | .8* | .1, 1.4 |
| Comorbid mental illness | –3.6** | –4.5, –2.7 | 3.1** | 2.3, 4.0 | .5** | .2, .9 |
| Alcohol dependence (reference: abuse) | –5.7** | –6.6, –4.8 | 5.0** | 4.2, 5.9 | .7** | .3, 1.0 |
| Health indicators | ||||||
| Diabetes | .4 | –1.3, 2.1 | −.1 | –1.6, 1.5 | −.3 | –1.2, .6 |
| Cirrhosis | –5.0 | –10.8, .8 | 6.1* | .3, 11.8 | –1.1** | –1.2, -.9 |
| Hepatitis B or C | –3.7 | –7.6, .1 | 1.1 | –1.4, 3.6 | 2.7 | −.2, 5.5 |
| Asthma | −.5 | –1.9, .9 | .6 | −.7, 1.9 | −.1 | −.7, .5 |
| HIV or AIDS | −.0 | –5.6, 5.6 | –1.9 | –4.6, .9 | 1.9 | –4.0, 7.8 |
| High blood pressure | −.4 | –1.8, 1.0 | .6 | −.7, 1.9 | −.2 | −.7, .3 |
*
p<0.05, **p<0.01.
Insurance status was also associated with trends in alcohol use treatment. Compared with uninsured individuals, those with Medicaid insurance had a lower probability of not receiving any treatment (marginal effect=−2.2 percentage points; 95% CI=−3.9 to −0.4). Moreover, Medicaid increased the probability of receiving treatment in a medical setting by 2.3 percentage points (95% CI=0.5–4.0).
Results from the logistic regression model for mental health treatment as the dependent variable were similar to those from the bivariate analysis (Table 4). After controlling for predisposing, enabling, and need factors, we found no significant differences in respondents’ likelihood of receiving any mental health treatment across the study period.
| Mental health treatment (baseline=23.3%)a | ||
|---|---|---|
| Variable | ME | 95% CI |
| Period (reference: 2008–2010) | ||
| 2011–2013 | −.6 | –1.8, .6 |
| 2014–2017 | –1.3 | –3.0, .4 |
| Male | –11.6** | –13.3, –9.8 |
| Age in years (reference: 18–25) | ||
| 26–34 | 4.3** | 2.4, 6.3 |
| 35–49 | 7.9** | 5.5, 10.2 |
| 50–64 | 8.1** | 4.9, 11.3 |
| Race-ethnicity (reference: non-Hispanic White) | ||
| Non-Hispanic Black | –8.7** | –10.5, –6.8 |
| Hispanic | –8.0** | –9.8, –6.1 |
| Non-Hispanic other | –7.4** | –9.3, –5.5 |
| Marital status (reference: married) | ||
| Widowed or divorced | 2.4* | .1, 4.7 |
| Never married | .6 | –1.4, 2.7 |
| Income in $ (reference: ≥75,000) | ||
| <20,000 | −.8 | –3.1, 1.6 |
| 20,000–49,999 | –2.3* | –4.1, −.6 |
| 50,000–74,999 | −.3 | –2.4, 1.9 |
| Self-reported health status (reference: excellent) | ||
| Very good | .1 | –1.9, 2.1 |
| Good | 3.0* | .7, 5.2 |
| Fair | 3.2* | .3, 6.1 |
| Poor | 3.1 | –2.9, 9.0 |
| Insurance status (reference: uninsured) | ||
| Private | 6.2** | 4.6, 7.8 |
| Medicaid | 10.8** | 7.8, 13.8 |
| Other insured | 9.7** | 5.4, 14.1 |
| Education (reference: high school graduate) | ||
| Less than high school | –2.7* | –5.0, -.5 |
| Some college | 1.9* | .1, 3.7 |
| College graduate and above | 6.0** | 3.6, 8.4 |
| Employment status (reference: full-time | ||
| Part-time | 2.8** | .7, 4.8 |
| Unemployed | 5.7** | 3.2, 8.2 |
| Other | 7.4** | 5.0, 9.8 |
| Comorbid substance use disorder | 3.3** | 1.5, 5.2 |
| Comorbid mental illness | 24.0** | 22.5, 25.6 |
| Alcohol dependence (reference: abuse) | 4.1** | 2.6, 5.7 |
| Health indicators | ||
| Diabetes | .8 | –2.9, 4.5 |
| Cirrhosis | 1.6 | –8.3, 11.5 |
| Hepatitis B or C | 5.9 | −.4, 12.3 |
| Asthma | 1.5 | −.5, 3.6 |
| HIV or AIDS | 13.9* | 1.8, 26.0 |
| High blood pressure | 3.9** | 1.7, 6.2 |
a
Sample of 36,628, compared with 36,707 for alcohol treatment.
*
p<0.05, **p<0.01.
Discussion
Treatment among individuals with alcohol use disorder declined from 6.9% in 2008–2010 to 5.6% in 2011–2013. When predisposing, enabling, and need-related factors were controlled for, significant decreases in alcohol use disorder treatment were observed in both medical and self-help settings. Moreover, during the study period, the probability of not receiving any treatment increased by 1.4 percentage points in 2011–2013 and 1.5 percentage points in 2014–2017 compared with the 2008–2010 baseline period.
Our results differ from those of previous studies of the trends in substance use disorder treatment. Earlier studies reported no increase in substance use or specialty substance use treatment among persons with a substance use disorder from 2005 to 2014 (14) or from 2011 to 2014 (16). The decline noted in this study may be due to the additional years of data available, differences in the population with an alcohol use disorder compared with the broader population with any substance use disorder, or both.
Although the proportion of those with alcohol use disorder who were uninsured declined during the study period, the probability of receiving no treatment increased. Several potential scenarios may explain the observed declines in treatment. First, even with expanded insurance coverage, the out-of-pocket cost for individuals seeking treatment may be high. Although MHPAEA was intended to reduce the treatment limitations and the cost-sharing patients face when seeking behavioral health care, recent research has put the effectiveness of MHPAEA into question. In 2017, substance use disorder care was more likely to be out of network than was medical or surgical care for inpatient visits (by 10.1 times), outpatient visits (8.5 times), and office visits (9.5 times) (27). These disparities increased for all service types from 2013 to 2017 (27). Out-of-network services have higher cost-sharing than in-network services and are therefore more expensive for patients. Although many individuals with alcohol use disorder gained insurance, they may have faced high out-of-pocket expenses for receiving alcohol treatment if they were seeking out-of-network services (28).
Another possible explanation for the decline in treatment for alcohol use disorder may be that the supply of alcohol treatment facilities became more limited during the study period, preventing individuals with alcohol use disorder from accessing treatment. However, data from the National Survey of Substance Abuse Treatment Services, an annual survey of substance use treatment facilities in the United States, indicate a modest increase in the number of facilities offering various alcohol use disorder treatments during the study period (29, 30). Even so, the number of individuals who received treatment for alcohol use disorder alone or for alcohol and drug use disorders in these facilities has decreased every year since 2011, declining from 756,890 in 2011 to 712,480 in 2017 (29). Future research should examine whether an increased emphasis on or increased demand for other types of substance use disorder treatment, such as opioid use disorder, has reduced these facilities’ capacity to treat patients with alcohol use disorder.
In addition to declines in overall treatment, self-help treatment declined by 0.5 percentage points in 2014–2017 compared with the 2008–2010 period. This finding is consistent with data indicating a steady decline in Alcoholics Anonymous members since 2001 (31). Future efforts to increase treatment rates may need to address attitudinal barriers, such as stigma and awareness of the effectiveness of therapies for alcohol use disorder and the financial barriers that were the focus of the ACA and MHPAEA.
The current findings suggest that expanded insurance was not enough to increase rates of alcohol use disorder treatment. Potential interventions to increase demand for treatment among individuals with alcohol use disorder may be one way to increase treatment. Public awareness campaigns in mass media could be a step in this direction, although they may need to be coupled with other demand-side and targeted interventions such as advertising restrictions and the targeting of high-risk groups, because the overall evidence base for their efficacy at increasing treatment uptake is relatively weak (32).
It is also notable that after we controlled for confounders in analyses, no significant changes were detected in the likelihood of receiving mental health treatment during the study period. Although the percentage of those with alcohol use disorder and a comorbid mental health condition significantly increased during the study period, we observed no commensurate increase in mental health treatment. These findings suggest that individuals with alcohol use disorder are not likely to substitute alcohol disorder treatment with mental health treatment.
This study had several limitations. Its analyses were descriptive. Other unmeasured constructs may have confounded changes over time, and causality could not be established in any reported associations. Thus, this study cannot provide the reasons for the unexpected decline in alcohol use disorder treatment observed for the study period. Several potential explanations need to be explored in future studies to better understand these trends, including changes in insurance benefits for alcohol use disorder treatment, changes in the geographic availability and capacity of substance use treatment facilities that offer alcohol use disorder services, changes in screening practices for alcohol use disorder, and changes in attitudes toward alcohol use disorder treatment. In addition, we used only 4 years of post-ACA data. Another limitation was that survey design changes in 2015 restricted the measurements of comorbid substance use disorders in these analyses to those consistently measured during the study period (26). Last, NSDUH excludes the institutionalized and noncivilian US population; the estimates from this study therefore cannot be generalized to those populations.
Conclusions
The findings of this study revealed unexpected declines during the 2008–2017 period in the treatment rates for alcohol use disorder in any setting, including medical and self-help settings, in both unadjusted and adjusted analyses. Given the gains in insurance coverage during this period, further research is needed to understand and address the factors responsible for the decline in treatment.
References
1.
Key Substance Use and Mental Health Indicators in the United States: Results From the 2018 National Survey on Drug Use and Health. Rockville, MD, Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality, 2019
2.
Rehm J: The risks associated with alcohol use and alcoholism. Alcohol Res Health 2011; 34:135–143
3.
Rehm J, Dawson D, Frick U, et al: Burden of disease associated with alcohol use disorders in the United States. Alcohol Clin Exp Res 2014; 38:1068–1077
4.
Esser MB, Sherk A, Liu Y, et al: Deaths and years of potential life lost from excessive alcohol use—United States, 2011-2015. MMWR Morb Mortal Wkly Rep 2020; 69:1428–1433
5.
Stahre M, Roeber J, Kanny D, et al: Contribution of excessive alcohol consumption to deaths and years of potential life lost in the United States. Prev Chronic Dis 2014; 11:E109
6.
Schmidt LA: Recent developments in alcohol services research on access to care. Alcohol Res 2016; 38:27–33
7.
Edlund MJ, Booth BM, Han X: Who seeks care where? Utilization of mental health and substance use disorder treatment in two national samples of individuals with alcohol use disorders. J Stud Alcohol Drugs 2012; 73:635–646
8.
Busch SH: Implications of the Mental Health Parity and Addiction Equity Act. Am J Psychiatry 2012; 169:1–3
9.
Barry CL, Huskamp HA: Moving beyond parity—mental health and addiction care under the ACA. N Engl J Med 2011; 365:973–975
10.
Mojtabai R, Feder KA, Kealhofer M, et al: State variations in Medicaid enrollment and utilization of substance use services: results from a National Longitudinal Study. J Subst Abuse Treat 2018; 89:75–86
11.
McLellan AT, Woodworth AM: The Affordable Care Act and treatment for “substance use disorders:” implications of ending segregated behavioral healthcare. J Subst Abuse Treat 2014; 46:541–545
12.
Tolbert J, Orgera K, Damico A: Key Facts About the Uninsured Population. Washington, DC, Kaiser Family Foundation, 2020. https://www.kff.org/uninsured/issue-brief/key-facts-about-the-uninsured-population. Accessed Nov 30, 2020
13.
Buck JA: The looming expansion and transformation of public substance abuse treatment under the Affordable Care Act. Health Aff 2011; 30:1402–1410
14.
Creedon TB, Cook BL: Access to mental health care increased but not for substance use, while disparities remain. Health Aff 2016; 35:1017–1021
15.
Saloner B, Lê Cook B: An ACA provision increased treatment for young adults with possible mental illnesses relative to comparison group. Health Aff 2014; 33:1425–1434
16.
Saloner B, Bandara S, Bachhuber M, et al: Insurance coverage and treatment use under the Affordable Care Act among adults with mental and substance use disorders. Psychiatr Serv 2017; 68:542–548
17.
Olfson M, Wall M, Barry CL, et al: Impact of Medicaid expansion on coverage and treatment of low-income adults with substance use disorders. Health Aff 2018; 37:1208–1215
18.
Olfson M, Wall M, Barry CL, et al: Effects of the Affordable Care Act on private insurance coverage and treatment of behavioral health conditions in young adults. Am J Public Health 2018; 108:1352–1354
19.
Feder KA, Mojtabai R, Krawczyk N, et al: Trends in insurance coverage and treatment among persons with opioid use disorders following the Affordable Care Act. Drug Alcohol Depend 2017; 179:271–274
20.
2017 National Survey on Drug Use and Health Final Analytic File Codebook. Rockville, MD, Substance Abuse and Mental Health Services Administration, Center for Behavioral Health Statistics and Quality, 2018
21.
Bandara SN, Samples H, Crum RM, et al: Is screening and intervention associated with treatment receipt among individuals with alcohol use disorder? Evidence from a national survey. J Subst Abuse Treat 2018; 92:85–90
22.
Johnson K, Rigg KK, Hopkins Eyles C: Receiving addiction treatment in the US: do patient demographics, drug of choice, or substance use disorder severity matter? Int J Drug Policy 2020; 75:102583
23.
Romo E, Ulbricht CM, Clark RE, et al: Correlates of specialty substance use treatment among adults with opioid use disorders. Addict Behav 2018; 86:96–103
24.
Olfson M, Wang S, Wall M, et al: Trends in serious psychological distress and outpatient mental health care of US adults. JAMA Psychiatry 2019; 76:152–161
25.
Andersen RM, Davidson PL: Improving access to care in America: individual and contextual indicators; in Changing the US Health Care System: Key Issues in Health Services Policy and Management, 3rd ed. Edited by Kominski GF. San Francisco, Jossey-Bass, 2007
26.
2015 National Survey on Drug Use and Health Public Use Data File Codebook. Rockville, MD, Substance Abuse and Mental Health Services Administration, 2018
27.
Davenport S, Gray TJ, Melek SP: Addiction and Mental Health vs Physical Health: Widening Disparities in Network Use and Provider Reimbursement. Seattle, Milliman, 2019
28.
Thomas CP, Hodgkin D, Levit K, et al: Growth in spending on substance use disorder treatment services for the privately insured population. Drug Alcohol Depend 2016; 160:143–150
29.
National Survey of Substance Abuse Treatment Services (N-SSATS): 2017, Data on Substance Abuse Treatment Facilities. Rockville, MD, Substance Abuse and Mental Health Services Administration, 2018
30.
National Survey of Substance Abuse Treatment Services (N-SSATS): 2011, Data on Substance Abuse Treatment Facilities. Rockville, MD, Substance Abuse and Mental Health Services Administration, 2012
31.
C J: AA Membership: Growth or Decline? AA Agnostica, 2015. https://aaagnostica.org/2015/09/06/aa-membership-growth-or-decline. Accessed March 6, 2019
32.
Young B, Lewis S, Katikireddi SV, et al: Effectiveness of mass media campaigns to reduce alcohol consumption and harm: a systematic review. Alcohol Alcohol 2018; 53:302–316
Information & Authors
Information
Published In
History
Received: 8 May 2020
Revision received: 13 November 2020
Revision received: 14 November 2021
Accepted: 21 December 2021
Published online: 23 February 2022
Published in print: September 01, 2022
Keywords
Authors
Funding Information
The authors report no financial relationships with commercial interests.
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
View Options
View options
PDF/EPUB
View PDF/EPUBLogin options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).